Printing defect detection method and system for polyester fabric

Abstract

The invention discloses a printing defect detection method and system for a polyester fabric, relates to the technical field of image detection, and aims to solve the problem that printing defects of the polyester fabric cannot be detected in the existing printing defect detection technology. The method solves the technical problems that the detection efficiency is low due to the fact that a convolution kernel has different sensitivities to different defect areas, the detection accuracy is influenced by image division areas, and the detection accuracy is reduced, and comprises the following steps: S1, obtaining a first production speed, determining a first target detection speed difference # imgabs0 # based on the first production speed and a polyester fabric printing production speed threshold value, and determining a second target detection speed difference # imgabs0 # based on the first target detection speed difference # imgabs0 #; a first detection position set is determined on the basis of the first target detection speed difference # imgabs1 #, a convolution process is constructed on the basis of the first detection position set, the target detection speed difference is determined according to the production speed by optimizing the detection principle and process, and the defect area is positioned, so that the detection efficiency is improved; through unique convolution and confidence coefficient screening, the detection precision is improved, and accurate and efficient detection is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of image detection, and more specifically, to a method and system for detecting printing defects of polyester fabrics. Background Art

[0002] In the field of polyester fabric printing production, due to the physical and chemical properties of polyester fabrics themselves, the defects produced during the printing process cannot be directly identified by the naked eye. Therefore, using machine vision technology to detect printing defects has become a necessary means in the industry.

[0003] The current mainstream machine vision algorithms mainly include the following clear steps: first, use the binarization technology based on RGB images to pre-process the collected images, the purpose of which is to convert the images into a form that is convenient for subsequent processing; then, build a convolution process based on the convolution kernel. In this process, the feature information in the image is extracted by sliding the convolution kernel on the image, and then these features are used to complete the model training; finally, use the trained model to perform defect detection on the real-time collected images.

[0004] However, the existing technology has a series of prominent problems: (1) The sensitivity of the convolution kernel to different defect areas varies significantly. This is because different types of printing defects have different features on the image. When extracting these features, the convolution kernel has a weaker response to the features of some defect areas, resulting in more time and computing resources required to detect these defects, thereby reducing detection efficiency.

[0005] (2) When the image is divided into regions based on transmission speed and computing performance to improve detection efficiency, if there are too many divided regions, the image details of each region will be lost, and the model will find it difficult to accurately extract defect features, which will lead to a decrease in detection accuracy.

[0006] Therefore, the traditional printing defect detection method for polyester fabrics only focuses on the recognition and processing of images, but does not propose effective solutions to the above-mentioned root problems affecting image quality. This greatly limits the accuracy and efficiency of printing defect detection for polyester fabrics, and becomes a bottleneck that needs to be broken through in the development of the industry. In view of this, we propose a printing defect detection method and system for polyester fabrics. Summary of the invention

[0007] The purpose of the present invention is to provide a method and system for detecting printing defects of polyester fabrics, so as to solve the technical problems in the existing polyester fabric printing defect detection technology, that is, the different sensitivities of the convolution kernel to different defect areas lead to low detection efficiency, the image division area affects the detection accuracy, and the detection accuracy is reduced.

[0008] To solve the above technical problems, the present invention provides the following technical solutions: A method for detecting printing defects of polyester fabrics, comprising the following steps: S1: Obtain the first production speed, determine the first target detection speed difference based on the first production speed and the polyester fabric printing production speed threshold , based on the first target detection speed difference determine the first detection position set, construct a convolution process based on the first detection position set, obtain the first convolution detection data segment set from the convolution process and the first detection position set, screen out the confidence detection data segment set based on the first convolution detection data segment set, and calculate the first confidence detection number based on the confidence detection data segment set ; S2: Construct the second production speed based on the first confidence detection number , determine the second target detection speed difference based on the second production speed , based on the second target detection speed difference determine the second detection position set, construct a convolution process based on the second detection position set, obtain the second convolution detection data segment set from the convolution process and the second detection position set, and screen out the confidence detection data segment set based on the second convolution detection data segment set; S3: Determine the detection result of printing defects of polyester fabrics based on the second confidence detection number ; Wherein, the step S1 includes the following sub-steps: S101: Obtain the first production speed, determine the first target detection speed difference based on the first production speed and the polyester fabric printing production speed threshold , the first target detection speed difference is determined based on the first production speed and the polyester fabric printing production speed threshold, and the specific formula is: ; Wherein, represents the difference between the polyester fabric printing production speed threshold and the first production speed, represents the first production speed, represents the distance that the polyester fabric moves per unit time on the printing equipment, represents the distance related to the movement of the polyester fabric , the first production speed and the polyester fabric printing production speed threshold corresponding function; S102: Determine the first detection position set based on the first target detection speed difference ; S103: Determine the confidence space set and the first convolution detection data segment set of the convolution process based on the first detection position set and the first production speed respectively; S104: Determine a confidence detection data segment set based on the first convolution detection data segment set and the confidence space set; S105: Calculate the first confidence detection number based on the confidence detection data segment set , if the first confidence detection number exceeds the preset first standard confidence detection number , then execute step S2.

[0009] Through optimization in terms of detection principle and process, when dealing with the sensitivity difference of convolution kernels to different defect regions, the present invention determines the first target detection speed difference by collecting the production speed in real time and combining it with the printing production speed threshold, accurately locates the regions prone to defects for key detection, reduces the ineffective detection range, and greatly improves the detection efficiency. At the same time, a unique convolution process is constructed to obtain the detection data segment set, and data is screened based on the confidence level, avoiding the problem of detail loss caused by image region division, enabling the model to accurately extract defect features, greatly improving the detection accuracy, and ensuring the accuracy and efficiency of polyester fabric printing defect detection.

[0010] Preferably, the step S102 includes: S102a: Determine a target position set based on the first standard deviation interval, where the first standard deviation interval is the speed deviation value of the polyester fabric without defects; S102b: Determine the first detection position set based on the target position set and the first target detection speed difference .

[0011] Preferably, in step S103, the method for determining the confidence space set of the convolution process includes the following steps: S103a: Obtain the target deviation time corresponding to the first target detection speed difference ; S103b: Determine the confidence space set based on the first detection position set and the target deviation time .

[0012] Preferably, the method for determining the confidence detection data segment set in the step S104 includes the following steps: S104a: Calculate the confidence vector set of the convolution process; S104b: Set the screening rule for the confidence detection data segment set based on the confidence vector set and the first convolution detection data segment set, and obtain the confidence detection data segment set from the screening rule for the confidence detection data segment set.

[0013] Preferably, the method for determining the value of the first standard deviation corresponding to the first target detection speed difference and the target position set includes the following steps: A: Calculate the target deviation time corresponding to the first detection position set, and multiply the target deviation time corresponding to the first detection position set by the target deviation time corresponding to the first detection position set to obtain the first time difference; B: The value of the first standard deviation satisfies the preset weight relationship of the first time difference, the second target deviation time, and the standard deviation of the polyester fabric. Let the value of the first standard deviation be , and the formula is: ; Among them, is the first target deviation time, is the second target deviation time, represents the error between the confidence space corresponding to the detection data segment and the confidence detection data segment, represents the first standard confidence deviation, represents the second standard confidence deviation; represents the first coefficient of the first standard deviation, represents the second coefficient of the second standard deviation.

[0014] Preferably, in the step S2, determining the detection result of the printing defect of the polyester fabric based on the second confidence detection number includes the following steps: S201: Based on the first confidence detection number construct the second production speed, and determine the second target detection speed difference based on the second production speed ; S202: Based on the second target detection speed difference determine the second detection position set, and construct the second convolutional detection data segment set based on the second detection position set and the first confidence detection data segment set; S203: Screen out the second confidence detection data segment set based on the second convolutional detection data segment set, and determine the second confidence detection number based on the second confidence detection data segment set ; S204: Determine the detection result of the printing defect of the polyester fabric based on the second confidence detection number . If the second confidence detection number of the detection result of the printing defect of the polyester fabric exceeds the second standard confidence detection number , then there are defects in the polyester fabric, otherwise there are no defects in the polyester fabric.

[0015] Preferably, the step S201 includes the following sub-steps: S201a: Based on the first confidence detection number construct the second production speed, and calculate the second coefficient corresponding to the first confidence detection number , the second production speed is ; S201b: Based on the second production speed and the polyester fabric printing production speed threshold, obtain the second target detection speed difference ; The second confidence detection number is calculated by the following method: , where is a constant coefficient.

[0016] Preferably, the step S202 includes the following sub-steps: S202a: Based on the second target detection speed difference determine the second detection position set; S202b: Based on the second detection position set and the second confidence detection data segment set, construct the second convolution detection data segment set.

[0017] Preferably, the step S203 includes the following sub-steps: S203a: Calculate the confidence vector set of the convolution process; S203b: Based on the confidence vector set and the second convolution detection data segment set, set the screening rule for the confidence detection data segment set, and obtain the confidence detection data segment set from the screening rule for the confidence detection data segment set.

[0018] A pattern defect detection system for polyester fabric, comprising: The first convolution detection data segment set determination module is used to obtain the first production speed, determine the first target detection speed difference based on the first production speed and the polyester fabric printing production speed threshold, determine the first detection position set based on the first target detection speed difference, construct the convolution process based on the first detection position set, and obtain the first convolution detection data segment set from the convolution process and the first detection position set; The second convolution detection data segment set determination module is used to screen out the confidence detection data segment set based on the first convolution detection data segment set, calculate the first confidence detection number based on the confidence detection data segment set, construct the second production speed based on the first confidence detection number, determine the second target detection speed difference based on the second production speed, determine the second detection position set based on the second target detection speed difference, construct the convolution process based on the second detection position set, and obtain the second convolution detection data segment set from the convolution process and the second detection position set; The detection result determination module is used to determine the polyester fabric printing defect detection result based on the first convolution detection data segment set and the second convolution detection data segment set.

[0019] Compared with the prior art, the beneficial effects of the present invention are: 1. By optimizing from the detection principle and process, when dealing with the sensitivity difference of convolution kernels to different defect regions, the present invention determines the first target detection speed difference by collecting the production speed in real time and combining it with the printing production speed threshold of polyester fabrics, accurately locates the regions prone to defects for key detection, reduces the ineffective detection range, and greatly improves the detection efficiency. At the same time, a unique convolution process is constructed to obtain a set of detection data segments, and the data is screened based on the confidence level, avoiding the problem of detail loss caused by image region division, enabling the model to accurately extract defect features, greatly improving the detection accuracy, and ensuring the accuracy and efficiency of the defect detection of polyester fabric printing.

[0020] 2. The present invention further improves the details of the detection process. During the construction of the convolution process, it deeply analyzes the relationship between the first target detection speed difference and various parameters, accurately determines the confidence space set, making the screened set of confidence detection data segments more representative. When calculating the confidence detection number, it comprehensively considers the influence of multiple parameters, deeply mines and analyzes the data, significantly improves the recognition ability of tiny and complex defects, and greatly enhances the reliability of the detection results.

[0021] 3. The present invention also constructs the second production speed based on the first confidence detection number, dynamically adjusts the subsequent detection parameters, enables the system to have self-adaptive ability, ensures the stable operation of the system, effectively reduces the equipment adaptation cost and maintenance cost of enterprises, and expands the practical application scope of the detection method and system. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 It is a schematic flow chart of the method of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0023] Example 1: As Figure 1 shown, a method for detecting printing defects of polyester fabrics according to the present invention includes the following steps: S1: Obtain the first production speed, determine the first target detection speed difference based on the first production speed and the printing production speed threshold of polyester fabrics , based on the first target detection speed difference determine the first detection position set, construct a convolution process based on the first detection position set, obtain the first convolution detection data segment set from the convolution process and the first detection position set, screen out the confidence detection data segment set based on the first convolution detection data segment set, and calculate the first confidence detection number based on the confidence detection data segment set ; In the embodiment of the present invention, the step S1 specifically includes the following sub-steps: S101: Obtain the first production speed, determine the first target detection speed difference based on the first production speed and the printing production speed threshold of polyester fabrics , the first production speed is The production speed of the polyester fabric printing equipment within a certain time, the first target detection speed difference Determined based on the first production speed and the polyester fabric printing production speed threshold, and the specific formula is: ; Wherein, Represents the first target detection speed difference, Represents the difference between the polyester fabric printing production speed threshold and the first production speed, Represents the first production speed, Represents the distance that the polyester fabric moves per unit time on the printing equipment, Represents the distance related to the movement of the polyester fabric , the first production speed And the polyester fabric printing production speed threshold Corresponding function; S102: Determine the first detection position set based on the first target detection speed difference ; In the embodiment of the present invention, the step S102 specifically includes: S102a: Determine the target position set based on the first standard deviation interval, and the first standard deviation interval is the speed deviation value of the polyester fabric when there is no defect; S102b: Determine the first detection position set based on the target position set and the first target detection speed difference ; S103: Determine the confidence space set and the first convolution detection data segment set of the convolution process respectively based on the first detection position set and the first production speed; In the embodiment of the present invention, in step S103, the method for determining the confidence space set of the convolution process specifically includes the following steps: S103a: Obtain the target deviation time Corresponding to the first target detection speed difference ; S103b: Determine the confidence space set based on the first detection position set and the target deviation time ; S104: Determine the confidence detection data segment set based on the first convolution detection data segment set and the confidence space set; In the embodiment of the present invention, the method for determining the confidence detection data segment set in the step S104 specifically includes the following steps: S104a: Calculate the confidence vector set of the convolution process; S104b: Based on the confidence vector set and the first convolutional detection data segment set, set the screening rules for the confidence detection data segment set, and obtain the confidence detection data segment set according to the screening rules for the confidence detection data segment set; S105: Based on the confidence detection data segment set, calculate the first confidence detection number ; In an embodiment of the present invention, the step S105 specifically includes the following steps: S105a: Based on the confidence detection data segment set and the convolutional process, calculate the first confidence detection number ; S105b: If the first confidence detection number exceeds the preset first standard confidence detection number , then execute step S2.

[0024] In an embodiment of the present invention, the value of the first standard deviation corresponding to the first target detection speed difference and the method for determining the target position set specifically includes the following steps: A: Calculate the target deviation time corresponding to the first detection position set, and multiply the target deviation time corresponding to the first detection position set by the target deviation time corresponding to the first detection position set to obtain the first time difference; B: The value of the first standard deviation satisfies the preset weight relationship of the first time difference, the second target deviation time, and the standard deviation of the polyester fabric. Let the value of the first standard deviation be , and the formula is: ; Among them, is the first target deviation time, is the second target deviation time, represents the error between the confidence space corresponding to the detection data segment and the confidence detection data segment, represents the first standard confidence deviation, represents the second standard confidence deviation; represents the first coefficient of the first standard deviation, represents the second coefficient of the second standard deviation; Among them, the larger, the greater the first target detection speed difference, , the larger, the faster the detection speed while ensuring the detection accuracy requirements.

[0025] By further improving the details of the detection process, during the construction of the convolution, deeply analyze the relationship between the first target detection speed difference and each parameter, accurately determine the confidence space set, and make the set of confidence detection data segments selected more representative. When calculating the confidence detection number, comprehensively consider the influence of multiple parameters, deeply mine and analyze the data, significantly improve the recognition ability of tiny and complex defects, and greatly enhance the reliability of the detection results.

[0026] S2: Based on the first confidence detection number Construct the second production speed, and determine the second target detection speed difference based on the second production speed , based on the second target detection speed difference Determine the second detection position set, construct the convolution process based on the second detection position set, obtain the second convolution detection data segment set from the convolution process and the second detection position set, and screen out the confidence detection data segment set based on the second convolution detection data segment set; Construct the second production speed based on the first confidence detection number, dynamically adjust the subsequent detection parameters, enable the system to have self-adaptive ability, ensure the stable operation of the system, effectively reduce the equipment adaptation cost and maintenance cost of the enterprise, and expand the practical application scope of the detection method and system; In the embodiment of the present invention, in step S2, based on the second confidence detection number Determining the detection result of the printing defect of the polyester fabric specifically includes the following steps: S201: Based on the first confidence detection number Construct the second production speed, and determine the second target detection speed difference based on the second production speed ; In the embodiment of the present invention, the step S201 specifically includes the following sub-steps: S201a: Based on the first confidence detection number Construct the second production speed, calculate the second coefficient corresponding to the first confidence detection number , and the second production speed is ; S201b: Based on the second production speed And the production speed threshold of the polyester fabric printing, obtain the second target detection speed difference ; The second confidence detection number Is calculated by the following method: , where Is a constant coefficient; S202: Based on the second target detection speed difference Determine the second detection position set, and construct the second convolution detection data segment set based on the second detection position set and the first confidence detection data segment set;​ In an embodiment of the present invention, the step S202 specifically includes the following sub-steps: S202a: Determine the second detection position set based on the second target detection speed difference Determine the second detection position set; S202b: Construct a second convolutional detection data segment set based on the second detection position set and the second confidence detection data segment set.

[0027] S203: Screen out the second confidence detection data segment set based on the second convolutional detection data segment set, and determine the second confidence detection number based on the second confidence detection data segment set ; In an embodiment of the present invention, the step S203 specifically includes the following sub-steps: S203a: Calculate the confidence vector set of the convolutional process; S203b: Based on the confidence vector set and the second convolutional detection data segment set, set the screening rule for the confidence detection data segment set, and obtain the confidence detection data segment set from the screening rule for the confidence detection data segment set; S204: Determine the detection result of the printing defect of the polyester fabric based on the second confidence detection number Determine the detection result of the printing defect of the polyester fabric; In an embodiment of the present invention, the method for determining the detection result of the printing defect of the polyester fabric in step S204 is as follows: If the second confidence detection number of the detection result of the printing defect of the polyester fabric exceeds the second standard confidence detection number , then the polyester fabric has a defect, otherwise the polyester fabric has no defect.

[0028] S3: Determine the detection result of the printing defect of the polyester fabric based on the second confidence detection number Determine the detection result of the printing defect of the polyester fabric.

[0029] By optimizing from the detection principle and process, when dealing with the sensitivity difference of the convolutional kernel to different defect regions, the present invention determines the first target detection speed difference by real-time collecting the production speed and combining the printing production speed threshold, accurately locates the regions prone to defects for key detection, reduces the ineffective detection range, and greatly improves the detection efficiency. At the same time, a unique convolutional process is constructed to obtain the detection data segment set, and the data is screened based on the confidence level, avoiding the problem of detail loss caused by image region division, enabling the model to accurately extract defect features, greatly improving the detection accuracy, and ensuring the accuracy and efficiency of the detection of the printing defects of the polyester fabric.

[0030] Embodiment 2: A pattern defect detection system for a polyester fabric, comprising: The first convolutional detection data segment set determination module is configured to obtain the first production speed, determine the first target detection speed difference based on the first production speed and the polyester fabric printing production speed threshold, determine the first detection position set based on the first target detection speed difference, construct a convolutional process based on the first detection position set, and obtain the first convolutional detection data segment set from the convolutional process and the first detection position set; The second convolutional detection data segment set determination module is configured to screen out the confidence detection data segment set based on the first convolutional detection data segment set, calculate the first confidence detection number based on the confidence detection data segment set, construct the second production speed based on the first confidence detection number, determine the second target detection speed difference based on the second production speed, determine the second detection position set based on the second target detection speed difference, construct a convolutional process based on the second detection position set, and obtain the second convolutional detection data segment set from the convolutional process and the second detection position set; The detection result determination module is configured to determine the polyester fabric printing defect detection result based on the first convolutional detection data segment set and the second convolutional detection data segment set.

[0031] The embodiments disclosed in the present invention are preferred embodiments, but not limited thereto. Those of ordinary skill in the art can easily understand the spirit of the present invention based on the above embodiments and make different extensions and changes. However, as long as they do not depart from the spirit of the present invention, they are within the protection scope of the present invention.

Claims

1. A method for detecting printing defects of polyester fabrics, characterized in that, Including the following steps: S1: Obtain the first production speed, determine the first target detection speed difference based on the first production speed and the polyester fabric printing production speed threshold, , based on the first target detection speed difference determine the first detection position set, construct a convolution process based on the first detection position set, obtain the first convolution detection data segment set from the convolution process and the first detection position set, filter out the confidence detection data segment set based on the first convolution detection data segment set, and calculate the first confidence detection number based on the confidence detection data segment set ; S2: Based on the first confidence detection number Construct a second production speed, and determine a second target detection speed difference based on the second production speed , based on the second target detection speed difference Determine a second set of detection positions, construct a convolution process based on the second set of detection positions, obtain a second set of convolution detection data segments from the convolution process and the second set of detection positions, and filter out a set of confidence detection data segments based on the second set of convolution detection data segments; S3: Based on the second confidence detection number Determine the detection result of printing defects on polyester fabrics; Among them, the step S1 includes the following sub-steps: S101: Obtain the first production speed, and determine the first target detection speed difference based on the first production speed and the production speed threshold for printing polyester fabrics , the first target detection speed difference is determined based on the first production speed and the production speed threshold for printing polyester fabrics. The specific formula is as follows: ； Among them, represents the difference between the printing production speed threshold of the polyester fabric and the first production speed, represents the first production speed, represents the distance that the polyester fabric moves per unit time on the printing equipment, represents the distance related to the movement of the polyester fabric , the first production speed and the printing production speed threshold of the polyester fabric corresponding function; S102: Based on the first target detection speed difference Determine the first detection position set; S103: Determine the confidence space set and the first convolution detection data segment set of the convolution process based on the first detection position set and the first production speed respectively; S104: Determine the confidence detection data segment set based on the first convolution detection data segment set and the confidence space set; S105: Calculate the first confidence detection number based on the set of confidence detection data segments , if the first confidence detection number exceeds the preset first standard confidence detection number , then execute step S2.

2. The printing defect detection method for a polyester fabric according to claim 1, characterized in that The step S102 includes: S102a: Determine the target position set based on the first standard deviation interval, and the first standard deviation interval is the speed deviation value of the polyester fabric when there is no defect; S102b: Based on the target position set and the first target detection speed difference Determine the first detection position set.

3. A method for detecting printing defects of a polyester fabric according to claim 2, characterized in that, In step S103, the method for determining the confidence space set of the convolution process includes the following steps: S103a: Obtain the first target detection speed difference The corresponding target deviation time ; S103b: Based on the first detection position set and the target deviation time Determine the confidence space set.

4. The printing defect detection method of a polyester fabric according to claim 3, characterized in that, The method for determining the confidence detection data segment set in step S104 includes the following steps: S104a: Calculate the confidence vector set of the convolution process; S104b: Based on the confidence vector set and the first convolution detection data segment set, set the screening rule for the confidence detection data segment set, and obtain the confidence detection data segment set from the screening rule of the confidence detection data segment set.

5. A method for detecting printing defects of a polyester fabric according to claim 4, characterized in that, The first target detection speed difference The method for determining the corresponding value of the first standard deviation and the set of target positions includes the following steps: A: Calculate the target deviation time corresponding to the first detection position set, and multiply the target deviation time corresponding to the first detection position set by the target deviation time corresponding to the first detection position set to obtain the first time difference; B: The value of the first standard deviation satisfies the preset weight relationship of the first time difference, the second target deviation time, and the standard deviation of the polyester fabric. Let the value of the first standard deviation be , and the formula is: ； Among them, is the first target deviation time, is the second target deviation time, represents the error between the confidence space corresponding to the detected data segment and the confidence detection data segment, represents the first standard confidence deviation, represents the second standard confidence deviation; represents the first coefficient of the first standard deviation, represents the second coefficient of the second standard deviation.

6. The printing defect detection method for a polyester fabric according to claim 1, characterized in that, In the step S2, based on the second confidence detection number Determining the detection result of the printing defect of the polyester fabric includes the following steps: S201: Based on the first confidence detection number Construct a second production speed, and determine a second target detection speed difference based on the second production speed ; S202: Based on the second target detection speed difference Determine the second detection position set, and construct a second convolutional detection data segment set based on the second detection position set and the first confidence detection data segment set; S203: Screen out the second confidence detection data segment set based on the second convolutional detection data segment set, and determine the second confidence detection number based on the second confidence detection data segment set ; S204: Based on the second confidence detection number Determine the detection result of the printing defect of the polyester fabric. If the second confidence detection number of the detection result of the printing defect of the polyester fabric exceeds the second standard confidence detection number , then there are defects in the polyester fabric, otherwise there are no defects in the polyester fabric.

7. A method for detecting printing defects of a polyester fabric according to claim 6, characterized in that, The step S201 includes the following sub-steps: S201a: Based on the first confidence detection number Construct the second production speed and calculate the first confidence detection number The corresponding second coefficient , and the second production speed is ; S201b: Based on the second production speed and the printing production speed threshold of polyester fabric, obtain the second target detection speed difference ; The second confidence detection number is calculated by the following method: , where is a constant coefficient.

8. A method for detecting printing defects of a polyester fabric according to claim 7, characterized in that, The step S202 includes the following sub-steps: S202a: Based on the second target detection speed difference Determine the second detection position set; S202b: Construct the second convolution detection data segment set based on the second detection position set and the second confidence detection data segment set.

9. The printing defect detection method for a polyester fabric according to claim 8, characterized in that, The step S203 includes the following sub-steps: S203a: Calculate the confidence vector set of the convolution process; S203b: Based on the confidence vector set and the second convolution detection data segment set, set the screening rule for the confidence detection data segment set, and obtain the confidence detection data segment set from the screening rule of the confidence detection data segment set.

10. A pattern defect detection system for polyester fabrics, which is implemented according to the printing defect detection method for polyester fabrics described in any one of claims 1-9, characterized in that Including: The first convolution detection data segment set determination module is used to obtain the first production speed, determine the first target detection speed difference based on the first production speed and the polyester fabric printing production speed threshold, determine the first detection position set based on the first target detection speed difference, construct the convolution process based on the first detection position set, and obtain the first convolution detection data segment set from the convolution process and the first detection position set; The second convolution detection data segment set determination module is used to screen out the confidence detection data segment set based on the first convolution detection data segment set, calculate the first confidence detection number based on the confidence detection data segment set, construct the second production speed based on the first confidence detection number, determine the second target detection speed difference based on the second production speed, determine the second detection position set based on the second target detection speed difference, construct the convolution process based on the second detection position set, and obtain the second convolution detection data segment set from the convolution process and the second detection position set; The detection result determination module is used to determine the polyester fabric printing defect detection result based on the first convolution detection data segment set and the second convolution detection data segment set.

Images